Radiation therapy has been used for many years to treat a patient by radiating selected anatomy so as to destroy tumors and the like.
Starting in the 1930's and continuing into the 1970's, “orthovoltage” X-ray therapy units of approximately 300 kVp/1.2 mm Cu HVL were widely used to treat malignancies (e.g., tumors). The word “orthovoltage” was based on the concept of “the right energy”, in the sense that a 300 kVp electrical potential, resulting in an approximately 160 keV X-ray beam, generates a photon energy such that the skin absorption and the bone absorption becomes approximately equal, which is a necessity for radiotherapy treatments within the skull, and an advantage for radiotherapy treatments elsewhere in the body. Energies lower than approximately 160 keV will deposit energy preferentially in bone. In this respect it should be appreciated that Compton scatter of the radiation within the bone effectively concentrates the radiation dose in the vasculature of the bone, which can destroy the bone vasculature. Thus, energies below approximately 160 keV are generally detrimental to the bone.
The 300 kVp units used into the 1970's had the disadvantage of excessive skin dose delivery (e.g., when compared to the radiation dose which is actually delivered to tumors in the body, after taking into account the absorption of the radiation by intervening tissue). This is particularly true when the radiotherapy is delivered using 1, 2 or 3 fixed entrance “portals”. In such a situation, the skin dose may be 3 to 5 times the tumor dose, and the skin is generally destroyed before the tumor is destroyed. Multiple fixed oversized entry portals, and multiple radiation sessions, were used as a partial solution to the problem of excessive skin dose delivery. However, it should be appreciated that the imaging of tumors at that time was comparatively crude, limited to perhaps 10 mm or so accuracy, and was severely hampered by the lack of precise knowledge of the tumor edges in 3D. As a result, practitioners generally resorted to stationary multiple entrance portals, multiple radiation sessions and generally oversized beams to deal with the uncertainty of tumor location. The locations of the entry portals, and their sizes, were generally determined by “best effort” projections based on 2D X-ray images. See Joseph Selman, MD, “Basic Physics of Radiation Therapy,” 1960, “Therapy Planning” pp 214-243.
It was recognized that higher photon energies could be used to address the known fall-off in radiation intensity that occurs with depth, whereby to improve the efficacy of the radiation treatment and spare the skin from the aforementioned radiation overdose problem.
To this end, efforts were made to create higher energy X-ray tubes. However, the engineering problems associated with creating direct X-ray tubes over 300-400 kVp were substantial. Several 1 MeV units were built, however, these high energy X-ray tubes generally used Van de Graaf generators (see, for example, the work of High Voltage Engineering and Massachusetts Institute of Technology) or RF resonant systems (see, for example, the Maxitron systems of General Electric), and were often 20 to 30 feet long.
The 1 to 2 MeV X-ray systems based on Van de Graaf generators and/or RF resonance beams generally fell out of favor due to the difficulty of moving the X-ray beams in 3D (e.g., because of the large physical size of the X-ray source enclosure).
In the mid-1950's, the “Gamma Knife” system (e.g., of Leksell) introduced the idea of using multiple isocentric beams, at higher energy levels, to provide radiotherapy to the patient. More particularly, the Gamma Knife system provided hundreds of high energy (approximate 1.2 MeV) spherically-arrayed cobalt beams focused to an isocenter. These beams were (and are) poorly defined due to the large size of the cobalt source (typically 10's of mm) and the short collimators which are inherent in the spherical array. This poor beam definition is problematic since sharp beam edges are necessary in order to avoid unintentionally radiating nearby healthy tissue. In addition, the Gamma Knife system provides stationary beams, and utilizes a fixed circular shape, thereby providing only spherical treatment volumes. Thus, the Gamma Knife system is unable to provide the variable, controllable treatment volumes necessary to closely match the shape of the anatomy which is to be treated. In addition, the Gamma Knife system suffers from diffuse scatter of the primary beam, which is inherent in energy levels over about 1.2 MeV.
The 1.2 MeV cobalt systems, including the Gamma Knife systems, generally fell out of favor largely due to their relatively large cobalt radiation sources (typically tens of mm) which caused the edges of the radiation beam to be very poorly defined, even with good collimators. This is due to the simple geometric problems inherent in collimating a large (e.g., 20 mm) radiation source (see FIG. 1) versus collimating a small (e.g., 1 mm) radiation source (see FIG. 2). More particularly, FIG. 1 shows a typically large cobalt source 5 of about 20 mm size, such as is commonly used in the Gamma Knife system, being collimated by a collimator 10. As can be seen from FIG. 1, it is not possible to collimate such a large source to a sharp edge (see “large error” 15), thereby effectively eliminating any possibility of providing a sharp clinical edge to the radiation beam. FIG. 2, on the other hand, is a schematic view showing the sharp edge resulting from a 1 mm X-ray source (or focal spot) 20 as it is collimated by a collimator 25, and the resulting “small error” 30. Note that with the large cobalt source 5, the penumbra of the beam edge may take 10 mm or more to “fade out” radially, from the nominal or desired edge, thereby making the beam emitted by a large cobalt source effectively useless if the desired treatment beam is to be 10 mm or less. This is a significant problem, given the current interest in small, accurate treatment volumes.
Subsequently, there was developed the linear particle accelerator (Linac) systems which are common today. These Linac systems typically have an energy level of about 6 MeV, which makes them highly effective in addressing the known fall-off in radiation intensity with depth, whereby to improve the efficacy of the radiation treatment and sparing the skin from radiation overdose. Although the secondary scatter envelope edge of a 6 MeV beam may be 4 to 5 times that of the early 300 kV beams (due primarily to in vitro tissue scatter, and not beam geometry), the secondary scatter problem was largely lost in the early days of Linacs due to the relatively low accuracy of tumor location which was common at the time. Note that in vitro tissue scatter of the radiation beam (which can reach 8-10 mm beyond the edge of a high energy radiation beam) cannot be suppressed by improved collimators.
Concurrent with the aforementioned improvements in radiotherapy systems, anatomical imaging systems have also been significantly improving, such that a 1 mm location accuracy is now routinely obtained. As a result, the X-ray beams can become smaller, and they can be more precisely positioned with the use of improved imaging modalities (e.g., CT, MRI and PET).
As a result, the tissue scattering which is inherent in high energy beams such as the current 6 MeV Linac systems, and which was previously of less concern due to the relatively low accuracy of tumor location, has now become a considerable problem.
More particularly, the secondary scatter effects of a high energy beam causes a “scatter” or “fog” of radiation that may have a range of up to 8-10 mm outside of the actual geometric beam. See FIG. 3, which shows the significant radiation scatter which can occur within tissue when using a high energy (e.g., 6 MeV) beam, and FIG. 4, which shows the nominal beam scatter which can occur within tissue when using a low energy (e.g., 100 kVp-800 kVp) beam. More particularly, FIG. 3 is a schematic view showing a typical 6 MeV Linac beam generated from an X-ray source 35, with secondary radiation 40 (generated by scatter interactions within the body) extending out to the dotted line 45. This is approximately 10 mm away from the geometric beam, generating the “scatter error” 50. Thus, even with excellent beam collimation, a 6 MeV beam is significantly scattered in tissue, again eliminating the achievement of a sharp edge. By comparison, FIG. 4 is a schematic view showing a 1.0 mm X-ray source 55, operating at 100-800 kVp so as to generate only short-range “Compton scatter” 60 and hence generating a sub-1 mm scatter error 65.
As noted above, the secondary scatter effects of a high energy (e.g., 6 MeV) beam did not present a significant problem with early Linac systems, inasmuch as tumor location was considerably less accurate than it is today, and inasmuch as the beams were fairly large and the gains in reducing excessive skin dose significantly outweighed “actual beam edge” definition drawbacks. It was accepted as the price of a large gain in efficacy and a significant reduction in skin damage. However, it is obviously a severe limit for the small, accurate beams desired in radiotherapy today. Others aware of these problems have recognized the medical physics at work here, and have outlined the potential benefits of “older” orthovoltage beams (e.g., 300 kVp beams) for higher net accuracy. See “Intermediate energy photons (1 MV) to improve dose gradient, conformality, and homogeneity: Potential benefits for small field intracranial radiosurgery”, Brian M. Keller, et al, Med. Phys. 36 (1), January 2009, © 2009 Am. Assoc. Phys. Med.; “Electron and photon spread contributions to the radiological penumbra for small monoenergetic x-ray beam (≤2 MeV), Jean-Philippe Pignol, et al, Journal of Applied Physics 105, 102011 (2009), © 2009 American Institute of Physics; and “Experimental measurement of radiological penumbra associated with intermediate energy x-rays (1 MV) and small radiosurgery field sizes”, Brian M. Keller, et al, Med Phys 34 (10), October 2007, © 2007 Am. Assoc. Phys. Med 3996.
A second major drawback associated with Linac systems, in particular for neuroradiosurgery, stems from the Linac's intended and historic advantage of high energy (and hence low fall-off as the beam travels within tissue). The high energy (e.g., 6 MeV) of the treatment beam causes the beam to continue on through the tissue at close to full destructive energy on the far side of the treatment zone, and the beam is increasing in diameter, and the “fog” of radiation still surrounds it, potentially wreaking havoc on normal tissue beyond or beside the treatment volume. This, combined with diffuse edges, pair production, and several other secondary effects associated with high energy beams, inherently limits what can be done to obtain a sharp beam edge at 6 MeV. By way of example but not limitation, with current Linac systems, even using multi-leaf collimators (2 to 3 mm are the smallest size that can work effectively) to tailor the 6 MeV beam, only 5 mm to 10 mm of beam precision is possible inside the body. This is ten times worse control than imaging accuracy would otherwise allow, and can prevent treatment of small lesions near critical anatomy such as the optic nerve.
Much of the recent prior art and current developments are geared to somehow gaining control of the Linac “fog of radiation” problems. See U.S. Pat. No. 8,280,001, issued Oct. 2, 2012 to Wang et al. From this observation, it is pointless to make a multi-leaf, “fine grain” collimation system, capable of a 1 mm geometric beam formation, if scatter within the patient will still be 5 to 8 times greater than the 1 mm geometric beam.
The presented X-ray edge definition problem is somewhat analogous to the situation where a very fine line (e.g., less than 1 mm) is to be drawn on paper—the pen used to draw it must be sharp (i.e., also less than 1 mm).